baoyu-markdown-to-html
Pass
Audited by Gen Agent Trust Hub on Jul 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious instructions, obfuscation, or unauthorized access patterns were identified in the skill's code or configuration.
- [COMMAND_EXECUTION]: The skill renders Mermaid diagrams to PNG images using a headless browser via the Chrome DevTools Protocol. This is a legitimate function for converting text-based diagrams into images.
- [EXTERNAL_DOWNLOADS]: The skill utilizes vendor-specific libraries and can fetch remote images referenced within the input Markdown files to facilitate correct HTML rendering.
- [DATA_EXFILTRATION]: The skill reads theme and preference configurations from the user's home directory (
$HOME/.baoyu-skills/). This is a standard practice for CLI-based tools and does not involve sensitive data harvesting. - [PROMPT_INJECTION]: The skill processes untrusted Markdown data, representing a surface for indirect prompt injection.
- Ingestion points: Markdown file content read from the file system in
scripts/main.ts. - Boundary markers: None explicitly defined for the input content; the skill relies on the
baoyu-mdlibrary for parsing. - Capability inventory: File system read/write (HTML generation, image caching) and network access (remote image fetching).
- Sanitization: Employs HTML attribute escaping for image tags to mitigate basic injection risks.
Audit Metadata